Nomad Places

Context & Problem Space

While working remotely in Brazil, I observed a recurring pain point in digital nomad communities. In every WhatsApp group, the same question surfaced daily: “Where’s a good place to work today?”.
From a product perspective, this was a clear signal of unmet user needs. The information existed in Google Maps reviews, but it was fragmented, unstructured, and hard to trust. Nomads needed a way to cut through noise and quickly assess the factors that mattered most: Wi-Fi quality, access to power outlets, and noise levels.

Discovery & Hypothesis

As in any product discovery phase, I started with user interviews and informal conversations. These validated that nomads didn’t want a glossy travel blog or influencer list, they needed utility, reliability, and speed. I defined my hypothesis: If I could aggregate reviews and extract the key insights that matter to nomads, then I could deliver a directory that saves them time and helps them choose work-friendly spots with confidence.

Constraints & Prioritization

This was a solo project, with limited time and resources. I had to prioritise:

• Do not manually visit or review each location.

• Instead, leverage automation and AI to handle the scale of reviews.

• Focus on MVP delivery rather than feature completeness.

This mirrors real-world product trade-offs: optimising for time-to-market and validated learning.

Solution Design & Tech Stack

Data Acquisition (Apify)
The first challenge was extracting reliable data at scale. Manually reading reviews wasn’t feasible, so I turned to Apify’s Google Maps Reviews Scraper. This allowed me to systematically collect reviews from hundreds of cafés and coworking spaces. As a PM, I treated this as solving the data ingestion layer of my product: ensuring I had a structured, repeatable pipeline of raw user feedback to build on.

Data Processing (Python)
Once the data was collected, it needed to be cleaned and aggregated. Using Python, I grouped reviews by location and consolidated them into single “voices” for each place. This was the data transformation step in the product lifecycle—turning messy, unstructured text into a manageable dataset that could be enriched and analyzed further.

AI Summarization (DeepSeek AI)
The core value came from AI. I used DeepSeek AI to process reviews and extract insights relevant to digital nomads—Wi-Fi, plugs, seating, and noise. This step acted as the intelligence layer, turning raw feedback into user-friendly narratives and structured ratings. It was like embedding a “review analyst” inside the product, delivering consistent summaries at scale.

Database & Structure (Airtable)
With AI outputs ready, I needed a flexible database to organize and manage the information. I chose Airtable for its ability to act as both a backend and lightweight CMS. It gave me relational structure, easy updates, and scalability—while also serving as a single source of truth for the directory’s content.

Frontend & Delivery (Softr)
Finally, I built the live directory in Softr, connected directly to Airtable. This allowed me to design a polished, user-facing product without custom coding. From a PM perspective, this was my delivery layer: where data and insights became a real, usable product that solved the community’s pain point.

Outcome & Impact

Within a few weeks, I had a working MVP: NomadPlaces.org, a searchable directory of cafés and coworking spaces enriched with AI-generated summaries and structured ratings.

Impact from a product lens:

• Time-to-market: Shipped an MVP in weeks, validating with early adopters in Rio and São Paulo.

• User value: Provided actionable insights (e.g., “Wi-Fi: Excellent, Noise: Mixed”) that directly answered the core JTBD (job to be done).

• Scalability: Designed a repeatable data pipeline that can scale to new cities with minimal effort.

Reflection & Learnings

From a product management perspective, NomadPlaces.org reinforced key principles:

• Problem-led building: The best ideas come from repeated user signals.

• AI as leverage: AI transforms messy, unstructured data into user-centric insights.

• No-code as speed: Pairing no-code with AI enables a solo PM to test and validate faster than traditional builds.

• Iteration focus: The MVP validated the need; future iterations could explore scaling in other cities and monetisation.

I had a lot of fun building NomadPlaces.org - it was a great exercise for me to learn new skills and to demonstrate my end-to-end product thinking, from identifying a pain point to delivering a working solution, using AI and no-code to accelerate the process.